Auto vs. Manual Campaigns: When to Switch (and Why Most Sellers Do It Too Early)

If you've read even a few Amazon PPC guides, you've probably seen this advice: "Start with auto campaigns for a week or two, then switch to manual." It's not wrong, exactly, but it skips the part that actually matters — auto and manual aren't a beginner and advanced version of the same thing. They do different jobs, and most sellers move to manual before they've actually gotten what auto is for.

What auto campaigns are actually for

Amazon's auto targeting doesn't just fill in for keyword research you haven't done yet. It's Amazon's own algorithm deciding, in real time, which search terms and products your listing is relevant to — using signals you don't have access to any other way. Every auto campaign you run is free market research: it tells you, in your own account's real data, what people actually type in before they buy your product. Not what a keyword tool guesses. What real shoppers really searched.

That data is the whole point of running auto campaigns at all. If you switch to manual after a week without ever reading the search term report, you've paid for the research and thrown away the results.

The real signal to switch — it's not a timer

"Run auto for two weeks" is a rule of thumb standing in for something more specific: run auto until you have enough orders to see a pattern, not until a certain number of days has passed. For a slow-moving listing, two weeks might get you three orders — not enough to tell you anything. For a fast mover, you might have a clear pattern in four or five days.

What you're actually watching for in the search term report is repetition: the same handful of search terms showing up across multiple orders, not just impressions. One converting search term is a data point. Three or four separate orders converting on the same or closely related terms is a pattern — that's your signal.

What "switching to manual" should actually mean

Here's the part that trips people up: switching to manual doesn't mean turning auto off. It means adding manual campaigns alongside auto, built from what auto just told you, while auto keeps running to find the next thing.

Concretely: - Pull the search terms from auto's report that converted more than once. - Build a manual campaign (or ad group) targeting those specific terms, so you can control the bid on each one individually instead of letting Amazon decide. - Add those same converting terms as negative exact matches on your auto campaign — so auto stops competing with your own manual campaign for a term you now have deliberately, and instead keeps spending its budget hunting for the next undiscovered term. - Leave auto running indefinitely. It's not a phase you graduate out of; it's your ongoing discovery engine, feeding new candidates into manual over time as your listing keeps getting new kinds of searches.

The actual mistake

The mistake isn't running manual campaigns too early in some vague sense — it's turning auto off once manual exists, or never cross-negating between the two. Do that, and you lose the discovery engine and start double-paying for terms both campaigns are quietly bidding on against each other.

Auto and manual aren't sequential stages. They're two different jobs — discovery and control — that should be running at the same time, for as long as your campaigns are live.

---

This is exactly the kind of decision AdArchitect makes for you automatically — which terms go where, when to cross-negate, and why — tailored to your actual product catalog. See how it works →